def main(extra_overplot=0):
overwrite = 1
groupname = 'NGC2354'
targetname = '554-051'
target_id = 5617126180115568256
cutoff_probability = 0.1
##########################################
# some data for a group that have a source_id list
cg18_path = os.path.join(
datadir, 'ngc_2354_CG18_subset.vot.gz')
group_df = given_votable_get_df(cg18_path, assert_equal='Source')
group_df = group_df[group_df['PMemb'] > cutoff_probability]
group_source_ids = np.array(group_df['Source']).astype(np.int64)
# todo: generalize above -- generically, it just needs some source id list.
##########################################
group_df_dr2 = given_source_ids_get_gaia_data(group_source_ids, groupname,
overwrite=overwrite)
target_d = objectid_search(target_id)
target_df = pd.read_csv(target_d['result'])
assert len(target_df) == 1
# now acquire the mean properties of the group, and query the neighborhood
# based on those properties. the number of neighbor stars to randomly
# select is min(5* the number of group members, 5000).
bounds = {}
params = ['parallax', 'ra', 'dec']
for param in params:
bounds[param+'_upper'] = (
group_df_dr2[param].mean() + 5*group_df_dr2[param].std()
)
bounds[param+'_lower'] = (
group_df_dr2[param].mean() - 5*group_df_dr2[param].std()
)
assert bounds['ra_upper'] < 360
assert bounds['ra_lower'] > 0
n_max = min((50*len(group_df_dr2), 10000))
nbhd_df = query_neighborhood(bounds, groupname, n_max=n_max,
overwrite=overwrite, is_cg18_group=True,
is_kc19_group=False, is_k13_group=False)
# ensure no overlap between the group members and the neighborhood sample.
common = group_df_dr2.merge(nbhd_df, on='source_id', how='inner')
snbhd_df = nbhd_df[~nbhd_df.source_id.isin(common.source_id)]
plot_group_neighborhood(targetname, groupname, group_df_dr2, target_df,
nbhd_df, cutoff_probability,
extra_overplot=extra_overplot)
def plot_TIC268_nbhd_small(outdir=RESULTSDIR):
set_style()
df = get_cdips_catalog(ver=0.4)
kc19_sel = ((df.cluster.str.contains('NGC_2516')) &
(df.reference.str.contains('Kounkel_2019')))
cg18_sel = ((df.cluster.str.contains('NGC_2516')) &
(df.reference.str.contains('CantatGaudin_2018')))
target_df = df[df.source_id == 5489726768531119616] # TIC 2683...
kc19_df = df[kc19_sel]
cg18_df = df[cg18_sel]
kc19_df = kc19_df[~(kc19_df.source_id.isin(cg18_df.source_id))]
##########
# NGC 2516 rough
bounds = {
'parallax_lower': 1.5,
'parallax_upper': 4.0,
'ra_lower': 108,
'ra_upper': 132,
'dec_lower': -76,
'dec_upper': -45
}
groupname = 'customngc2516'
nbhd_df = query_neighborhood(bounds,
groupname,
n_max=6000,
overwrite=False,
manual_gmag_limit=17)
sel_nbhd = ((~nbhd_df.source_id.isin(kc19_df.source_id))
& (~nbhd_df.source_id.isin(cg18_df.source_id)))
from copy import deepcopy
orig_nbhd_df = deepcopy(nbhd_df)
nbhd_df = nbhd_df[sel_nbhd]
print(f'Got {len(nbhd_df)} neighbors')
print(f'Got {len(cg18_df)} in core')
print(f'Got {len(kc19_df)} in corona')
##########
plt.close('all')
f, axs = plt.subplots(figsize=(4, 3), ncols=2)
xv, yv = 'ra', 'dec'
axs[0].scatter(nbhd_df[xv],
nbhd_df[yv],
c='gray',
alpha=0.5,
zorder=2,
s=7,
rasterized=True,
linewidths=0,
label='Field',
marker='.')
axs[0].scatter(kc19_df[xv],
kc19_df[yv],
c='lightskyblue',
alpha=0.9,
zorder=3,
s=7,
rasterized=True,
linewidths=0.15,
label='Corona',
marker='.',
edgecolors='k')
axs[0].scatter(cg18_df[xv],
cg18_df[yv],
c='k',
alpha=0.9,
zorder=4,
s=7,
rasterized=True,
label='Core',
marker='.')
axs[0].plot(target_df[xv],
target_df[yv],
alpha=1,
mew=0.5,
zorder=8,
label='TOI 1937',
markerfacecolor='lightskyblue',
markersize=14,
marker='*',
color='black',
lw=0)
axs[0].set_xlabel(r'$\alpha$ [deg]')
axs[0].set_ylabel(r'$\delta$ [deg]')
axs[0].set_xlim([108, 132])
axs[0].set_ylim([-76, -45])
##########
get_yval = (lambda _df: np.array(_df['phot_g_mean_mag'] + 5 * np.log10(_df[
'parallax'] / 1e3) + 5))
get_xval = (lambda _df: np.array(_df['phot_bp_mean_mag'] - _df[
'phot_rp_mean_mag']))
axs[1].scatter(get_xval(nbhd_df),
get_yval(nbhd_df),
c='gray',
alpha=0.8,
zorder=2,
s=7,
rasterized=True,
linewidths=0,
label='Field',
marker='.')
axs[1].scatter(get_xval(kc19_df),
get_yval(kc19_df),
c='lightskyblue',
alpha=1,
zorder=3,
s=7,
rasterized=True,
linewidths=0.15,
label='Corona',
marker='.',
edgecolors='k')
axs[1].scatter(get_xval(cg18_df),
get_yval(cg18_df),
c='k',
alpha=0.9,
zorder=4,
s=7,
rasterized=True,
linewidths=0,
label='Core',
marker='.')
axs[1].plot(get_xval(target_df),
get_yval(target_df),
alpha=1,
mew=0.5,
zorder=8,
label='TOI 1937',
markerfacecolor='lightskyblue',
markersize=14,
marker='*',
color='black',
lw=0)
axs[1].set_ylim(axs[1].get_ylim()[::-1])
axs[1].set_xlabel('Bp - Rp [mag]')
axs[1].set_ylabel('Absolute G [mag]', labelpad=-6)
##########
words = ['Field', 'Corona', 'Core', 'TOI1937'][::-1]
colors = ['gray', 'lightskyblue', 'k', 'lightskyblue'][::-1]
rainbow_text(0.98, 0.02, words, colors, size='medium', ax=axs[0])
f.tight_layout(w_pad=2)
outpath = os.path.join(outdir, 'small_ngc2516.png')
savefig(f, outpath)
def get_neighborhood_information(
source_id,
overwrite=0,
min_n_nbhrs=1000,
manual_gmag_limit=None
):
"""
Given a source_id for a star (potentially a field star), acquire
information necessary for neighborhood diagnostic plots.
Parameters:
source_id: Gaia DR2 source_id
overwrite: Whether the Gaia cache gets overwritten.
manual_gmag_limit: G < manual_gmag_limit for the neighborhood
"""
#
# Get the targetname
#
ticid = gaiadr2_to_tic(str(source_id))
toiid = ticid_to_toiid(ticid)
if isinstance(toiid, str):
targetname = toiid
else:
targetname = 'TIC{}.01'.format(ticid)
#
# Get Gaia information for target.
#
enforce_all_sourceids_viable = True
savstr = '_nbhdonly'
target_d = objectid_search(
source_id,
columns=('source_id', 'ra','dec', 'ra_error', 'dec_error',
'phot_g_mean_mag', 'phot_bp_mean_mag', 'phot_rp_mean_mag',
'l','b', 'parallax, parallax_error', 'pmra','pmra_error',
'pmdec','pmdec_error', 'radial_velocity'),
forcefetch=True,
gaia_mirror='vizier'
)
target_df = pd.read_csv(target_d['result'])
assert len(target_df) == 1
# now acquire the mean properties of the group, and query the neighborhood
# based on those properties. the number of neighbor stars to randomly
# select is min(5* the number of group members, 5000). (cutoff group
# bounds based on parallax because further groups more uncertain).
bounds = {}
params = ['parallax', 'ra', 'dec']
plx_mean = float(target_df.parallax)
n_nbhrs = 0
n_std = 5
n_std_incr = 10
n_std_max = 200
if plx_mean > 10:
n_std = 5
n_std_incr = 20
n_std_max = 1000
while n_nbhrs < min_n_nbhrs:
if n_std > n_std_max:
return None
LOGINFO('trying when bounding by {} stdevns'.format(n_std))
for param in params:
mult = 1 if 'parallax' in param else 2
bounds[param+'_upper'] = (
float(target_df[param]) + mult*n_std*float(target_df[param + '_error'])
)
bounds[param+'_lower'] = (
float(target_df[param]) - mult*n_std*float(target_df[param + '_error'])
)
if bounds['ra_upper'] > 360:
bounds['ra_upper'] = 359.99
if bounds['ra_lower'] < 0:
bounds['ra_lower'] = 0
if bounds['parallax_lower'] < 0:
bounds['parallax_lower'] = 0
n_max = int(1e4)
if manual_gmag_limit is None:
manual_gmag_limit = 17
groupname = '{}'.format(source_id)
# only force overwrite if iterating
if n_nbhrs == 0:
nbhd_df = query_neighborhood(bounds, groupname, n_max=n_max,
overwrite=overwrite,
manual_gmag_limit=manual_gmag_limit)
else:
nbhd_df = query_neighborhood(bounds, groupname, n_max=n_max,
overwrite=True,
manual_gmag_limit=manual_gmag_limit)
n_nbhrs = len(nbhd_df)
LOGINFO(42*'=')
LOGINFO('Got {} neighborhods, when minimum was {}'.
format(n_nbhrs, min_n_nbhrs))
LOGINFO(42*'=')
n_std += n_std_incr
n_std = 3
pmdec_min = np.nanmean(nbhd_df['pmdec']) - n_std*np.nanstd(nbhd_df['pmdec'])
pmdec_max = np.nanmean(nbhd_df['pmdec']) + n_std*np.nanstd(nbhd_df['pmdec'])
pmra_min = np.nanmean(nbhd_df['pmra']) - n_std*np.nanstd(nbhd_df['pmra'])
pmra_max = np.nanmean(nbhd_df['pmra']) + n_std*np.nanstd(nbhd_df['pmra'])
pmdec_min = min((pmdec_min, float(target_df['pmdec'])))
pmdec_max = max((pmdec_max, float(target_df['pmdec'])))
pmra_min = min((pmra_min, float(target_df['pmra'])))
pmra_max = max((pmra_max, float(target_df['pmra'])))
return (targetname, groupname, target_df, nbhd_df,
pmdec_min, pmdec_max, pmra_min, pmra_max)
def get_group_and_neighborhood_information(
source_id,
overwrite=0,
force_groupname=None,
force_references=None,
force_cdips_match=True,
manual_gmag_limit=None,
CATALOG_VERSION=0.6):
"""
Given a source_id for a cluster member, acquire information necessary for
neighborhood diagnostic plots. (Namely, find all the group members, then do
a Gaia query of everything).
Parameters:
source_id: Gaia DR2 source_id
overwrite: Whether the Gaia cache gets overwritten.
Optional kwargs: force_groupname and force_references. If passed, for
example as the arrays ["kc19group_1222"] and ["Kounkel_2019"], plot
generation will be forced without verifying that the target "source_id"
is a member of the group.
"""
if not isinstance(source_id, np.int64):
source_id = np.int64(source_id)
cdips_df = get_cdips_pub_catalog(ver=CATALOG_VERSION)
row = cdips_df[cdips_df.source_id == source_id]
if pd.isnull(row['cluster'].iloc[0]):
LOGWARNING(f'Did not find any group matches for GAIA DR2 {source_id}.')
return None
#
# Get numpy arrays of references and cluster names for this star.
# This is needed b/c multiple memberships are comma-separated.
#
if 'reference' in row:
references = np.array(row['reference'].iloc[0].split(','))
elif 'reference_id' in row:
references = np.array(row['reference_id'].iloc[0].split(','))
clusters = np.array(row['cluster'].iloc[0].split(','))
if force_references and force_groupname:
references = force_references
clusters = np.array([force_groupname])
assert len(references) == len(clusters)
if row is None and force_cdips_match:
LOGINFO('Failed to get CDIPS target list match for {}'.format(source_id))
return None
#
# Given the numpy array of clusters, find the best cluster membership list
# to make the report with.
#
from cdips.catalogbuild.membership_lists import RANKED_MEMBERSHIP_DICT
references_in_ensembles = RANKED_MEMBERSHIP_DICT[CATALOG_VERSION]['isgroup']
references_in_field = RANKED_MEMBERSHIP_DICT[CATALOG_VERSION]['isfield']
is_in_group = np.any(np.in1d(references, references_in_ensembles))
is_in_field = np.any(np.in1d(references, references_in_field))
if is_in_group:
# At least one of the references given corresponds to an actual coeval
# group, and not a list of young field stars. In this case, take the
# highest precedence group (lowest index) as the one to make the plot
# for.
referencename = references_in_ensembles[
int(min(np.argwhere(np.in1d(references_in_ensembles, references))))
]
groupname = clusters[references == referencename][0]
else:
LOGWARNING(f'Did not find any group matches for GAIA DR2 {source_id}.')
return None
#
# Get the targetname
#
ticid = gaiadr2_to_tic(str(source_id))
toiid = ticid_to_toiid(ticid)
if isinstance(toiid, str):
targetname = toiid
else:
targetname = 'TIC{}.01'.format(ticid)
#
# Get the group members!
#
# (We avoided the field star case earlier.)
cdips_df = cdips_df[~pd.isnull(cdips_df.cluster)]
group_df = cdips_df[
# avoids e.g., "kc19group_981" matching on "kc19group_98". deals with
# cases beginning of string, middle of string, and end of string.
(
(cdips_df.cluster.str.contains(groupname+','))
|
(cdips_df.cluster.str.contains(','+groupname+','))
|
(cdips_df.cluster.str.contains(','+groupname+'$'))
|
(cdips_df.cluster == groupname)
)
&
(cdips_df.reference_id.str.contains(referencename))
]
group_source_ids = np.array(group_df['source_id']).astype(np.int64)
np.testing.assert_array_equal(group_df['source_id'], group_source_ids)
#
# Given the source ids, get all the relevant Gaia information.
#
enforce_all_sourceids_viable = True
savstr = f'_{groupname.replace(" ","_")}_{referencename}'
group_df_dr2 = given_source_ids_get_gaia_data(
group_source_ids, groupname.replace(" ","_"), overwrite=overwrite,
enforce_all_sourceids_viable=enforce_all_sourceids_viable,
n_max=min((len(group_source_ids), 10000)),
savstr=savstr
)
target_d = objectid_search(
source_id,
columns=('source_id', 'ra','dec', 'phot_g_mean_mag',
'phot_bp_mean_mag', 'phot_rp_mean_mag', 'l','b',
'parallax, parallax_error', 'pmra','pmra_error',
'pmdec','pmdec_error', 'radial_velocity'),
forcefetch=True,
gaia_mirror='vizier'
)
target_df = pd.read_csv(target_d['result'])
assert len(target_df) == 1
# now acquire the mean properties of the group, and query the neighborhood
# based on those properties. the number of neighbor stars to randomly
# select is min(5* the number of group members, 5000). (cutoff group
# bounds based on parallax because further groups more uncertain).
bounds = {}
params = ['parallax', 'ra', 'dec']
plx_mean = group_df_dr2['parallax'].mean()
if plx_mean > 5:
n_std = 5
elif plx_mean > 3:
n_std = 4
else:
n_std = 3
LOGINFO(f'bounding by {n_std} stdevns')
for param in params:
bounds[param+'_upper'] = np.minimum(
group_df_dr2[param].mean() + n_std*group_df_dr2[param].std(),
360-1e-5
)
bounds[param+'_lower'] = np.maximum(
group_df_dr2[param].mean() - n_std*group_df_dr2[param].std(),
1e-5
)
if bounds['parallax_lower'] < 0:
bounds['parallax_lower'] = 0
assert bounds['ra_upper'] < 360
assert bounds['ra_lower'] > 0
assert bounds['parallax_lower'] >= 0
n_max = min((50*len(group_df_dr2), 10000))
if manual_gmag_limit is None:
manual_gmag_limit = np.nanpercentile(group_df_dr2.phot_g_mean_mag,95)
mstr = savstr
nbhd_df = query_neighborhood(bounds, groupname.replace(" ","_"), n_max=n_max,
overwrite=overwrite,
manual_gmag_limit=manual_gmag_limit,
mstr=mstr)
# ensure no overlap between the group members and the neighborhood sample.
common = group_df_dr2.merge(nbhd_df, on='source_id', how='inner')
snbhd_df = nbhd_df[~nbhd_df.source_id.isin(common.source_id)]
n_std = 5
pmdec_min = group_df_dr2['pmdec'].mean() - n_std*group_df_dr2['pmdec'].std()
pmdec_max = group_df_dr2['pmdec'].mean() + n_std*group_df_dr2['pmdec'].std()
pmra_min = group_df_dr2['pmra'].mean() - n_std*group_df_dr2['pmra'].std()
pmra_max = group_df_dr2['pmra'].mean() + n_std*group_df_dr2['pmra'].std()
pmdec_min = min((pmdec_min, float(target_df['pmdec'])))
pmdec_max = max((pmdec_max, float(target_df['pmdec'])))
pmra_min = min((pmra_min, float(target_df['pmra'])))
pmra_max = max((pmra_max, float(target_df['pmra'])))
return (targetname, groupname, referencename, group_df_dr2, target_df,
snbhd_df, pmdec_min, pmdec_max, pmra_min, pmra_max)
def _get_fullfaint_edr3_dataframes():
"""
Return: nbhd_df, core_df, halo_df, full_df, target_df
(for NGC 2516, "full faint" sample -- i.e., as faint as possible, but
***after crossmatching the GAIA DR2 targets with GAIA EDR3***. This
crossmatch is run using the dr2_neighbourhood table from the Gaia archive,
and then taking the closest angular separation match for cases with
multiple matches.)
Further notes are in "_get_fullfaint_dataframes" docstring.
This procedure yields:
FOR DR2:
Got 1106 in fullfaint CG18
Got 3003 in fullfaint KC19
Got 1860 in fullfaint M21
Got 1912 in fullfaint KC19 after removing core matches
Got 1096 in fullfaint M21 after removing core matches
Got 280 in fullfaint M21 after removing KC19 matches
Got 13834 neighbors
Got 1106 in core
Got 2192 in corona
Got 1091 KC19 / CG18 overlaps
Got 764 M21 / CG18 overlaps
FOR EDR3:
Got 1106 EDR3 matches in core.
99th pct [arcsec] 1577.8 -> 0.3
Got 1912 EDR3 matches in KC19.
99th pct [arcsec] 1702.8 -> 0.5
Got 280 EDR3 matches in M21.
99th pct [arcsec] 1426.6 -> 0.3
Got 13843 EDR3 matches in nbhd.
99th pct [arcsec] 1833.9 -> 3.7
(((
CG18/core: got 1143 matches vs 1106 source id queries.
KC19/halo: got 2005 matches vs 1912 source id queries
Nbhd: got 15123 matches vs 13843 source id queries.
)))
"""
# get the full CG18 NGC 2516 memberships, downloaded from Vizier
cg18path = os.path.join(DATADIR, 'gaia',
'CantatGaudin2018_vizier_only_NGC2516.fits')
hdul = fits.open(cg18path)
cg18_tab = Table(hdul[1].data)
cg18_df = cg18_tab.to_pandas()
cg18_df['source_id'] = cg18_df['Source']
# get the full KC19 NGC 2516 memberships, from Marina's file
# NGC 2516 == "Theia 613" in Kounkel's approach.
kc19path = os.path.join(DATADIR, 'gaia', 'string_table1.csv')
kc19_df = pd.read_csv(kc19path)
kc19_df = kc19_df[kc19_df.group_id == 613]
# get the full M21 NGC 2516 memberships
m21path = os.path.join(DATADIR, 'gaia',
'Meingast_2021_NGC2516_all1860members.fits')
m21_df = Table(fits.open(m21path)[1].data).to_pandas()
m21_df = m21_df.rename(mapper={'GaiaDR2': 'source_id'}, axis=1)
print(42 * '=' + '\nFOR DR2:')
print(f'Got {len(cg18_df)} in fullfaint CG18')
print(f'Got {len(kc19_df)} in fullfaint KC19')
print(f'Got {len(m21_df)} in fullfaint M21')
kc19_cg18_overlap_df = kc19_df[(kc19_df.source_id.isin(cg18_df.source_id))]
kc19_df = kc19_df[~(kc19_df.source_id.isin(cg18_df.source_id))]
print(f'Got {len(kc19_df)} in fullfaint KC19 after removing core matches')
m21_cg18_overlap_df = m21_df[(m21_df.source_id.isin(cg18_df.source_id))]
m21_df = m21_df[~(m21_df.source_id.isin(cg18_df.source_id))]
print(f'Got {len(m21_df)} in fullfaint M21 after removing core matches')
m21_df = m21_df[~(m21_df.source_id.isin(kc19_df.source_id))]
print(f'Got {len(m21_df)} in fullfaint M21 after removing KC19 matches')
##########
# NGC 2516 rough
bounds = {
'parallax_lower': 1.5,
'parallax_upper': 4.0,
'ra_lower': 108,
'ra_upper': 132,
'dec_lower': -76,
'dec_upper': -45
}
groupname = 'customngc2516_fullfaint'
nbhd_df = query_neighborhood(bounds,
groupname,
n_max=14000,
overwrite=False,
manual_gmag_limit=19)
sel_nbhd = ((~nbhd_df.source_id.isin(kc19_df.source_id))
& (~nbhd_df.source_id.isin(cg18_df.source_id))
& (~nbhd_df.source_id.isin(m21_df.source_id)))
orig_nbhd_df = deepcopy(nbhd_df)
nbhd_df = nbhd_df[sel_nbhd]
print(f'Got {len(nbhd_df)} neighbors')
print(f'Got {len(cg18_df)} in core')
print(f'Got {len(kc19_df)+len(m21_df)} in corona')
print(f'Got {len(kc19_cg18_overlap_df)} KC19 / CG18 overlaps')
print(f'Got {len(m21_cg18_overlap_df)} M21 / CG18 overlaps')
assert (len(cg18_df) + len(kc19_df) + len(m21_df) == len(
np.unique(np.array(pd.concat(
(cg18_df, kc19_df, m21_df))['source_id']))))
cg18_df_edr3 = (given_dr2_sourceids_get_edr3_xmatch(
nparr(cg18_df.Source).astype(np.int64),
'fullfaint_ngc2516_cg18_df',
overwrite=False))
kc19_df_edr3 = (given_dr2_sourceids_get_edr3_xmatch(
nparr(kc19_df.source_id).astype(np.int64),
'fullfaint_ngc2516_kc19_df',
overwrite=False))
m21_df_edr3 = (given_dr2_sourceids_get_edr3_xmatch(
nparr(m21_df.source_id).astype(np.int64),
'fullfaint_ngc2516_m21_df',
overwrite=False))
nbhd_df_edr3 = (given_dr2_sourceids_get_edr3_xmatch(
nparr(nbhd_df.source_id).astype(np.int64),
'fullfaint_ngc2516_nbhd_df',
overwrite=False))
print(42 * '=' + '\nFOR EDR3:')
# Take the closest (proper motion and epoch-corrected) angular distance as
# THE single match.
get_edr3_xm = lambda _df: (_df.sort_values(by='angular_distance').
drop_duplicates(subset='dr2_source_id',
keep='first'))
s_cg18_df_edr3 = get_edr3_xm(cg18_df_edr3)
s_kc19_df_edr3 = get_edr3_xm(kc19_df_edr3)
s_m21_df_edr3 = get_edr3_xm(m21_df_edr3)
s_nbhd_df_edr3 = get_edr3_xm(nbhd_df_edr3)
print(
f'Got {len(s_cg18_df_edr3)} EDR3 matches in core.\n' +
f'99th pct [arcsec] {np.nanpercentile(cg18_df_edr3.angular_distance, 99):.1f} -> {np.nanpercentile(s_cg18_df_edr3.angular_distance, 99):.1f}'
)
print(
f'Got {len(s_kc19_df_edr3)} EDR3 matches in KC19.\n' +
f'99th pct [arcsec] {np.nanpercentile(kc19_df_edr3.angular_distance, 99):.1f} -> {np.nanpercentile(s_kc19_df_edr3.angular_distance, 99):.1f}'
)
print(
f'Got {len(s_m21_df_edr3)} EDR3 matches in M21.\n' +
f'99th pct [arcsec] {np.nanpercentile(m21_df_edr3.angular_distance, 99):.1f} -> {np.nanpercentile(s_m21_df_edr3.angular_distance, 99):.1f}'
)
print(
f'Got {len(s_nbhd_df_edr3)} EDR3 matches in nbhd.\n' +
f'99th pct [arcsec] {np.nanpercentile(nbhd_df_edr3.angular_distance, 99):.1f} -> {np.nanpercentile(s_nbhd_df_edr3.angular_distance, 99):.1f}'
)
# Finally, query Gaia EDR3 to get the latest and greatest fullfaint
# photometry
kc19_df_0 = given_source_ids_get_gaia_data(
np.array(s_kc19_df_edr3.dr3_source_id),
'fullfaint_ngc2516_kc19_df_edr3',
n_max=10000,
overwrite=False,
enforce_all_sourceids_viable=True,
gaia_datarelease='gaiaedr3')
cg18_df_0 = given_source_ids_get_gaia_data(
np.array(s_cg18_df_edr3.dr3_source_id),
'fullfaint_ngc2516_cg18_df_edr3',
n_max=10000,
overwrite=False,
enforce_all_sourceids_viable=True,
gaia_datarelease='gaiaedr3')
m21_df_0 = given_source_ids_get_gaia_data(
np.array(s_m21_df_edr3.dr3_source_id),
'fullfaint_ngc2516_m21_df_edr3',
n_max=10000,
overwrite=False,
enforce_all_sourceids_viable=True,
gaia_datarelease='gaiaedr3')
nbhd_df_0 = given_source_ids_get_gaia_data(
np.array(s_nbhd_df_edr3.dr3_source_id),
'fullfaint_ngc2516_nbhd_df_edr3',
n_max=15000,
overwrite=False,
enforce_all_sourceids_viable=True,
gaia_datarelease='gaiaedr3')
assert len(cg18_df) == len(cg18_df_0)
assert len(kc19_df) == len(kc19_df_0)
assert len(m21_df) == len(m21_df_0)
assert len(nbhd_df) == len(nbhd_df_0)
# nb. these "source_ids" are now EDR3 source_ids.
np.testing.assert_array_equal(np.array(kc19_df_0.source_id),
np.array(kc19_df_0.source_id_2))
np.testing.assert_array_equal(np.array(cg18_df_0.source_id),
np.array(cg18_df_0.source_id_2))
np.testing.assert_array_equal(np.array(m21_df_0.source_id),
np.array(m21_df_0.source_id_2))
np.testing.assert_array_equal(np.array(nbhd_df_0.source_id),
np.array(nbhd_df_0.source_id_2))
kc19_df_0['dr2_source_id'] = nparr(s_kc19_df_edr3['dr2_source_id']).astype(
np.int64)
cg18_df_0['dr2_source_id'] = nparr(s_cg18_df_edr3['dr2_source_id']).astype(
np.int64)
m21_df_0['dr2_source_id'] = nparr(s_m21_df_edr3['dr2_source_id']).astype(
np.int64)
nbhd_df_0['dr2_source_id'] = nparr(s_nbhd_df_edr3['dr2_source_id']).astype(
np.int64)
target_df = kc19_df_0[kc19_df_0.source_id ==
5489726768531119616] # TIC 2683...
#
# wrap up into the full source list
#
cg18_df_0['subcluster'] = 'core'
kc19_df_0['subcluster'] = 'halo'
m21_df_0['subcluster'] = 'halo'
core_df = cg18_df_0
halo_df = pd.concat((kc19_df_0, m21_df_0)).reset_index()
full_df = pd.concat((core_df, halo_df)).reset_index()
assert len(np.unique(full_df.source_id)) == len(full_df)
print(f'Got {len(full_df)} unique sources in the cluster.')
full_df['in_CG18'] = full_df.source_id.isin(cg18_df.source_id)
full_df['in_KC19'] = full_df.source_id.isin(kc19_df.source_id)
full_df['in_M21'] = full_df.source_id.isin(m21_df.source_id)
nbhd_df['dr2_radial_velocity'] = nbhd_df['radial_velocity']
return nbhd_df, core_df, halo_df, full_df, target_df
def _get_fullfaint_dataframes():
"""
Return: nbhd_df, core_df, halo_df, full_df, target_df
(for NGC 2516, "full faint" sample -- i.e., as faint as possible.)
The "core" is all available Cantat-Gaudin 2018 members, with no magnitude
cutoff.
The "halo" is the full Kounkel & Covey 2019 + Meingast 2021 member set,
provided that the source is not in the core. (i.e., KC19 and M21 get no
points for getting the "core" targets correct).
The "neighborhood" was selected via
bounds = { 'parallax_lower': 1.5, 'parallax_upper': 4.0, 'ra_lower': 108,
'ra_upper': 132, 'dec_lower': -76, 'dec_upper': -45 }
nbhd_df = query_neighborhood(bounds, groupname, n_max=14000,
overwrite=False, manual_gmag_limit=19)
This procedure yields:
Got 1106 in fullfaint CG18
Got 3003 in fullfaint KC19
Got 1860 in fullfaint M21
Got 1912 in fullfaint KC19 after removing core matches
Got 1096 in fullfaint M21 after removing core matches
Got 280 in fullfaint M21 after removing KC19 matches
Got 13834 neighbors
Got 1106 in core
Got 2192 in corona
Got 1091 KC19 / CG18 overlaps
Got 764 M21 / CG18 overlaps
Got 3298 unique sources in the cluster.
"""
# get the full CG18 NGC 2516 memberships, downloaded from Vizier
cg18path = os.path.join(DATADIR, 'gaia',
'CantatGaudin2018_vizier_only_NGC2516.fits')
hdul = fits.open(cg18path)
cg18_tab = Table(hdul[1].data)
cg18_df = cg18_tab.to_pandas()
cg18_df['source_id'] = cg18_df['Source']
# get the full KC19 NGC 2516 memberships, from Marina's file
# NGC 2516 == "Theia 613" in Kounkel's approach.
kc19path = os.path.join(DATADIR, 'gaia', 'string_table1.csv')
kc19_df = pd.read_csv(kc19path)
kc19_df = kc19_df[kc19_df.group_id == 613]
# get the full M21 NGC 2516 memberships
m21path = os.path.join(DATADIR, 'gaia',
'Meingast_2021_NGC2516_all1860members.fits')
m21_df = Table(fits.open(m21path)[1].data).to_pandas()
m21_df = m21_df.rename(mapper={'GaiaDR2': 'source_id'}, axis=1)
print(f'Got {len(cg18_df)} in fullfaint CG18')
print(f'Got {len(kc19_df)} in fullfaint KC19')
print(f'Got {len(m21_df)} in fullfaint M21')
kc19_cg18_overlap_df = kc19_df[(kc19_df.source_id.isin(cg18_df.source_id))]
kc19_df = kc19_df[~(kc19_df.source_id.isin(cg18_df.source_id))]
print(f'Got {len(kc19_df)} in fullfaint KC19 after removing core matches')
m21_cg18_overlap_df = m21_df[(m21_df.source_id.isin(cg18_df.source_id))]
m21_df = m21_df[~(m21_df.source_id.isin(cg18_df.source_id))]
print(f'Got {len(m21_df)} in fullfaint M21 after removing core matches')
m21_df = m21_df[~(m21_df.source_id.isin(kc19_df.source_id))]
print(f'Got {len(m21_df)} in fullfaint M21 after removing KC19 matches')
##########
# NGC 2516 rough
bounds = {
'parallax_lower': 1.5,
'parallax_upper': 4.0,
'ra_lower': 108,
'ra_upper': 132,
'dec_lower': -76,
'dec_upper': -45
}
groupname = 'customngc2516_fullfaint'
nbhd_df = query_neighborhood(bounds,
groupname,
n_max=14000,
overwrite=False,
manual_gmag_limit=19)
# query gaia DR2 to get the fullfaint photometry
kc19_df_0 = given_source_ids_get_gaia_data(
np.array(kc19_df.source_id),
'ngc2516_kc19_earhart_fullfaint',
n_max=10000,
overwrite=False,
enforce_all_sourceids_viable=True)
cg18_df_0 = given_source_ids_get_gaia_data(
np.array(cg18_df.Source),
'ngc2516_cg18_earhart_fullfaint',
n_max=10000,
overwrite=False,
enforce_all_sourceids_viable=True)
m21_df_0 = given_source_ids_get_gaia_data(
np.array(m21_df.source_id),
'ngc2516_m21_earhart_fullfaint',
n_max=10000,
overwrite=False,
enforce_all_sourceids_viable=True)
assert len(cg18_df) == len(cg18_df_0)
assert len(kc19_df) == len(kc19_df_0)
assert len(m21_df) == len(m21_df_0)
target_df = kc19_df_0[kc19_df_0.source_id ==
5489726768531119616] # TIC 2683...
sel_nbhd = ((~nbhd_df.source_id.isin(kc19_df.source_id))
& (~nbhd_df.source_id.isin(cg18_df.source_id))
& (~nbhd_df.source_id.isin(m21_df.source_id)))
orig_nbhd_df = deepcopy(nbhd_df)
nbhd_df = nbhd_df[sel_nbhd]
print(f'Got {len(nbhd_df)} neighbors')
print(f'Got {len(cg18_df)} in core')
print(f'Got {len(kc19_df)+len(m21_df)} in corona')
print(f'Got {len(kc19_cg18_overlap_df)} KC19 / CG18 overlaps')
print(f'Got {len(m21_cg18_overlap_df)} M21 / CG18 overlaps')
#
# wrap up into the full source list
#
cg18_df_0['subcluster'] = 'core'
kc19_df_0['subcluster'] = 'halo'
m21_df_0['subcluster'] = 'halo'
core_df = cg18_df_0
halo_df = pd.concat((kc19_df_0, m21_df_0)).reset_index()
full_df = pd.concat((core_df, halo_df)).reset_index()
assert len(np.unique(full_df.source_id)) == len(full_df)
print(f'Got {len(full_df)} unique sources in the cluster.')
full_df['in_CG18'] = full_df.source_id.isin(cg18_df.source_id)
kc19_df = pd.read_csv(kc19path)
kc19_df = kc19_df[kc19_df.group_id == 613]
full_df['in_KC19'] = full_df.source_id.isin(kc19_df.source_id)
m21_df = Table(fits.open(m21path)[1].data).to_pandas()
m21_df = m21_df.rename(mapper={'GaiaDR2': 'source_id'}, axis=1)
full_df['in_M21'] = full_df.source_id.isin(m21_df.source_id)
return nbhd_df, core_df, halo_df, full_df, target_df